西北旱区遥感影像分类的支持向量机法  

The support vector machine method for RS images' classification in northwest arid area

作　　者：张静[1] 张翔[1] 田龙[1] 张青峰[1] 

机构地区：[1]西北农林科技大学资源环境学院,陕西杨凌712100

出　　处：《测绘科学》2017年第1期49-52,58,共5页Science of Surveying and Mapping

基　　金：国家科技支撑计划子课题项目(2011BAD29B09-1-1C);中央高校基本科研业务费专项资金项目(QN2012042);西北农林科技大学博士科研启动基金项目(201104050395)

摘　　要：针对较大范围、不同时相、不同气候和地貌类型的遥感影像的土地利用现状分类问题,提出了一种结合标准植被指数和纹理特征的支持向量机法。此方法改进了陕西延安、甘肃嘉峪关和青海果洛的遥感影像分类,有效地解决了最大似然法和BP神经网络法的缺陷造成的分类精度不高的问题。分类结果表明:与最大似然法和BP神经网络法相比,结合标准植被指数和纹理特征的支持向量机法的分类总精度最高(97.75%),Kappa系数为0.9691。该方法可为西北旱区遥感影像解译和土地资源可持续发展战略提供方法支撑。Aiming at the classification method for the field of land resources of remote sensing images,which cover a large area with different climate and geomorphology at different phases,this paper provides a support vector machine（SVM）method based on the NDVI and texture feature.It improve the classification precision of the northwest arid areas at Yanan in Shaanxi,Jiayuguan in Gansu and Guoluo in Qinghai,solved the problem of not high classification accuracy in the maximum likelihood and BP neural network method because of their own defects.The result shows that：compared with the maximum likelihood method and BP neural network method,the SVM method has the highest classification accuracy（97.75%）with its Kappa coefficient was 0.9691.The SVM method can provide a methodological reference for the land resources sustainable development strategic.

关 键 词：遥感应用 西北旱区 支持向量机 遥感影像分类 

分 类 号：P208[天文地球—地图制图学与地理信息工程]